What would you think of a course in: l French literature, based entirely on a Reader's Digest English version of The Three Musketeers

l geography that still taught continental drift, without mentioning plate tectonics.

l physics of Newton's Laws that did not mention that these are an approximation - because the teacher had not heard of relativity.

l chemistry that omitted pH as too complex for students to understand, but suggested that dipping litmus paper in gives a good enough answer.

É«ºÐÖ±²¥

But this is the attitude of many university staff towards data analysis and modern information technology. I was recently asked to suggest software that was "simple to use'' for students who would do an introductory statistics course to include "the chi-squared test''. I knew perfectly well that what was intended was the conventional test of independence on a two-way table, but to make the point I asked which chi-squared test. The lecturer looked blank. Unfortunately, this would be the reaction from almost any department where a "short statistics course'' is pushed on to first and second-year students.

The point highlights a reluctance to update and take account of IT. To adhere to teaching " the chi-squared test on 2 x 2 tables'' is equivalent to any of the examples listed. It preserves the attitude that statistics is a cookbook of abstruse calculations carried out by researchers as a formality before publication. It ignores almost 40 years of developments in techniques and numerical methods made feasible by the use of computers for processing data.

É«ºÐÖ±²¥

Gavin Fairbairn (Multimedia, February 10) comments on the attitude that word processing is purely a tool for producing "nice, neat work''. I agree that, on the contrary, students must be taught skills in laying out work (rather than being presented with an interface of extraordinary complexity and invited to play); they must then be guided in using the tool for the craft of writing. Similarly, I would argue that the information society must be based on teaching new skills in handling information. It is axiomatic when applying computers that the user must understand the task being performed. Assuming that the computer will deliver sensible answers with the minimum control or understanding on the part of the user leads only to disappointment.

What is wrong with much of the activity in "teaching with IT" is that teachers expect to carry on exactly as before. They hope desperately that adding computer-based teaching will enable the system to teach faster, cheaper, and with less effort on the part of the students.

Old learning was based on specialists who had spent years learning their subject, the accepted facts, the sources of information and the manners of thinking. The trouble with that era was that it led to specialists, to different methods and ways of thinking in different disciplines. Information technology transcends the barriers; through the Internet, anyone, anywhere can have access to the latest information and ideas. It could be a release from ever-more specialisation.

There is no shortage of visionaries, at national and local levels. My university is working on an information strategy. Academic visions can resemble military strategies; they ignore reality and leave it to subordinates to bridge the gap between technology and the information nirvana. The Follett report has the example of Alice in her undergraduate cell calling up the works of Gunter Grass and Heinrich Boll; what a shame if her search is frustrated by the German index not recognising that she wants "Gunter" and "Boll". This is not an unrealistic or even rare problem; I have just searched for "Coto Donana" in a large database and found only one hit - but my keyboard does not offer direct input of Do$ana. Over the past year I have addressed several groups of email users on the problems of the non-English character set.

É«ºÐÖ±²¥

The potential is great, but it does require a proper syllabus in information science as a new literacy. Some aspects are very basic. If using a word processor obviates the use of a typist, the author must acquire the basic skills to produce an at least "adequate" typescript. At a higher level, the writer should learn how to arrange and express ideas, to analyse data with current techniques, to incorporate network material without laboriously retyping it, to recognise where human intelligence and computer complement each other.

Someone has to identify and analyse the tasks for which IT is suited. Someone should also produce a believable cost/benefit statement. I am as convinced as anyone that we are moving towards an information society, but cannot go along with pious assumptions that human-computer interactions are painless, intuitive and educational. In the cases of academic writing and statistical data analysis, computers give us abilities that were only dreams 20 years ago. Current students must be trained in current techniques, and have their imaginations fired to accept the next 20 years. It is up to their teachers to keep one step ahead.

R. Allan Reese Head of applications, Hull University Computer Centre email: r.a.reese@ucc.hull.ac.uk